By David W. Body

March 8, 2018

Executive Summary

The Affordable Care Act (ACA) contains an insidious feature that can cost (or save) people thousands of dollars based on whether their household income is greater (or less) than a specific amount. People purchasing health insurance through the ACA qualify for premium subsidies if their income is less than 400% of the federal poverty level (FPL). These subsidies can be substantial, even tens of thousands of dollars per year. But if household income exceeds 400% of the FPL, even by a dollar, no subsidy is available, and any subsidies already received must be repaid. This subsidy “cliff” amounts to a an effective marginal income tax rate in excess of 100% and can create a disincentive to work for those affected. This document describes how subsidies and the subsidy “cliff” are calculated and looks at how the size of the cliff varies with age and geographically.

Main features of the ACA

The Affordable Care Act (ACA) was enacted in 2010 and major provisions went into effect in 2014. The goal, to expand health insurance coverage in the United States, was accomplished in two primary ways.

First, Medicaid eligibility was expanded to all U.S. citizens and legal residents with incomes up to 133% of the federal poverty level. As of October 2017, 32 states (including DC) had expanded Medicaid and 19 had not.

Second, significant changes were made to individual health insurance markets.

The law specifies four tiers of coverage: bronze, silver, gold, and platinum, which are expected to cover 60%, 70%, 80%, or 90%, respectively, of an enrollee’s health care costs.

Two kinds of subsidies

There are two kinds of subsidies under the ACA.

Premium tax credits

The larger of the two kinds of subsidies is used to help people pay premiums of policies purchased on health insurance exchanges set up under the ACA. These subsidies are provided in the form of “premium tax credits” and are available to ACA enrollees earning up to 400% of the federal poverty level (FPL). Premium tax credits are calculated based on an enrollee’s expected income and can be paid in advance directly to insurance companies so enrollees can pay lower monthly premiums. When enrollees file their income tax returns the following year, premium tax credits are reconciled based on actual income, and income tax liability or refund is adjusted accordingly.

Cost sharing reductions

Cost sharing reduction (CSR) subsidies are provided to reduce the copayments and deductibles for ACA enrollees with incomes under 250% of the FPL. Until 2018, these subsidies were paid to insurance companies to compensate for the costs of reducing copayments and deductibles for eligible ACA enrollees. Starting in 2018, insurance companies no longer receive the CSR subsidies, but are still required to reduce copayments and deductibles. Insurance companies raised 2018 premiums to compensate for the loss of the CSR subsidies.

As we’ll see below, these premium increases and the resulting increase in premium tax credits creates perverse effects for certain individuals.

How is the premium tax credit calculated?

The larger of the two subsidies is the premium tax credit which is available to people whose income is up to 400% of the FPL. This subsidy is designed to make health insurance cost no more than a certain percentage of a household’s income. The subsidy is calculated as follows:

For 2018, the applicable percentage is determined from the following table, based on household income:

Household income percentage of FPL Initial percentage Final percentage
Less than 133% 2.01% 2.01%
At least 133% but less than 150% 3.02% 4.03%
At least 150% but less than 200% 4.03% 6.34%
At least 200% but less than 250% 6.34% 8.10%
At least 250% but less than 300% 8.10% 9.50%
At least 300% but less than 400% 9.56% 9.56%

Subsidies are available to people who purchase health insurance on an exchange and whose household income is between 100% and 400% of the FPL. For people in states that have expanded Medicaid under the ACA, subsidies are available for incomes between 138% and 400% of FPL (because almost everyone with an income below 138% of FPL qualifies for Medicaid and therefore is not eligible to buy insurance on an exchange).

The applicable percentages look like this:

(As we’ll see later, the discrete jump in applicable percentage at 133% of FPL for non-Medicaid expansion states results in a mini-cliff at that income level.)

The maximum available subsidy will be the excess of the reference premium over the applicable percentage of a household’s income.

Example

Here is a concrete example of how the premium tax credit is calculated.

Assume a married couple both age 30 with no children and a household income of $44,660 living in Polk County, Iowa. The 2018 FPL for an Iowa household with 2 people is $16,240. Their income is therefore 275% of the FPL, which means the applicable percentage from the table above will be half-way between 8.10% and 9.50%, or 8.8%. The reference premium is the second lowest silver premium available. For this couple in 2018, only one silver plan is available, so the reference premium is the premium for that plan, which is $1,165.22 per month, or $13,982.64 per year. The maximum available subsidy will be the amount by which the reference premium exceeds 8.8% of their income, or $13,982.64 - 0.088 x $44,660, which equals $10,052.56.

Regardless of whether this family purchases a gold policy for $15,551.04 per year, or a silver policy for $13,982.64 per year, or a bronze policy for $11,342.88 per year, their subsidy will be the full $10,052.56 because this is less than the cost of any of these policies.

Subsidies rise with premiums

Note that subsidy amounts are driven directly by premiums. Since the available subsidy is the excess of the reference premium over a percent of a household’s income, the available subsidy rises dollar for dollar with the reference premium, all other things being equal. High premiums therefore translate directly into high subsidies.

Subsidies decrease as income rises

As household income rises, the maximum available subsidy is reduced both by the fact that the applicable percentage applies to a higher income, and by the fact that the applicable percentage gradually rises to 9.56%. This effectively raises the marginal income tax rate for the household, because additional income results in loss of premium tax credits.

It is critical to understand that subsidies are only available to people whose income does not exceed 400% of the FPL. Subsidies are not available when household income exceeds 400% of FPL. The complete loss of subsidies once household income reaches 400% of FPL creates a “cliff” and implies an effective marginal income tax rate far in excess of 100%. For a household right at 400% of FPL, a single dollar of additional income will result in a subsidy loss of potentially thousands of dollars. Any subsidies that have already been received will be repaid in the form of an additional liability on the federal income tax return filed the following year.

The following chart shows the maximum available subsidy as a function of household income for the household in the example above.

As shown in the chart, there is a subsidy “cliff” at 400% of FPL, or $64,960. In this example, the size of the cliff is $7,772. So, for example, if this family receives an advanced premium tax credit based on an estimated income of less than $64,960, but their actual income for 2018 is even one dollar over this amount, they will be liable to repay the entire amount of their subsidy when they file their 2018 federal income tax return in 2019.

In a given year, the size of the cliff is mainly a function of the second highest silver premium available, which varies by location, household size, and age. Health insurance premiums generally increase with the age of the insureds, so older people tend to face higher subsidy cliffs.

Here are the same charts for 2018 for an individual age 21, and couples age 40, 50, and 60 in Polk County, Iowa.

For comparison, here are the same charts for Douglas County, Nebraska. Nebraska did not expand Medicaid, which accounts for the mini-cliff at 133% of FPL.

The 2018 cliff amount for a couple age 60 in Des Moines (Polk County), IA is $27,225. In Omaha (Douglass County), NE it’s a whopping $32,126. As we’ll see below, the 2018 cliff amount can go even higher.

The cliff is high when premiums are high

Why are the cliff amounts so high? The cliff amount is simply the available subsidy at 400% of FPL. High premiums translate directly into high subsidies which translate directly into high cliff amounts.

Variation in cliff amounts

We’ve already seen that cliff amounts rise with age because premiums (and therefore subsidies) rise with age, and that they also vary with location.

To get an idea of how cliff amounts vary by location, let’s calculate the size of the cliff for similar households in every county in the US. (Only states that use the federal health insurance exchange are included. States that operate their own exchanges are not included.)

Here are summary statistics for the 2018 cliff amounts across all counties in the federal health insurance exchange for couples age 30, 40, 50, and 60 with no children.

2018 ACA subsidy cliff amounts for all counties
Age minimum q1 median mean q3 maximum
30 0 2748 4691 4775 6181 16138
40 364 3876 6064 6149 7742 18953
50 2977 7886 10943 11072 13289 28956
60 7750 15210 19855 20026 23378 47228

The following histograms show that distributions for all age examples are skewed, with cliff amounts generally higher for older households.

Geographic variation

The following maps show how the size of the cliff varies geographically. As mentioned earlier, data is only included for states that use the federal health insurance exchange. States that run their own exchanges are coded as NA. The size of the cliff varies with location and shows strong geographical patterns. These patterns appear very similar regardless of age.

It is interesting that some entire states have high or low cliffs more or less statewide. For example, Alaska, Iowa, Nebraska, Tennessee, Utah, and Wyoming appear to have high cliff amounts statewide, while Hawaii, Indiana, Michigan, Montana, New Hampshire, New Mexico, North Dakota, and Oregon appear to have low cliff amounts statewide. On the other hand, several states show much more variation from county to county, for example Arizona, Florida, Georgia, Maine, Nevada, Pennsylvania, Texas, Virginia, and Wisconsin. We know the cliff amount is primarily a function of the reference silver premium for each county, but we don’t have an explanation for why these premiums vary like they do.

Summary

The Affordable Care Act requires insurers to offer insurance to everyone on the same terms, meaning they cannot discriminate based on pre-existing conditions. To avoid adverse selection, almost everyone is required to have a certain level of health insurance coverage or pay a penalty. Subsidies are provided to make the required coverage affordable. These subsidies are based on household income and phase out as income rises. However, no subsidy is available if household income exceeds 400% of the federal poverty level (FPL), and any subsidy received must be repaid. High premiums translate directly into high subsidies, which in turn translate into a high “cliff” at 400% of FPL.

The size of the cliff varies with age and geographically, and can easily be tens of thousands of dollars for older households in high-premium locations. Households that are affected therefore stand to lose tens of thousands of dollars if their income exceeds 400% of FPL, which amounts to a marginal income tax rate in excess of 100% and creates a substantial disincentive to work.

Questions for further research

Who is affected?

Who are the people most affected by high ACA subsidy cliffs and how many are there? We can speculate that some of them are likely self-employed small business owners or farmers who don’t have access to group health insurance. Older households face higher premiums and therefore higher cliff amounts.

How do affected people respond?

A self-employed person facing a high subsidy cliff may be able to limit their income to avoid going over the cliff, or they may decided self-employment is no longer viable and seek employment with health insurance benefits. Others may decided to forgo health insurance and pay a penalty, while also facing the prospect of paying for health care without insurance. (The penalty has been repealed starting in 2019.) Data about household responses to the high subsidy cliffs in 2018 won’t be available for a while, but it would be interesting to find out how people actually do respond. It would also be interesting to find out how many people are unaware they are affected.

How has the size of the cliff changed over time?

We know higher premiums translate into higher cliffs. We also know that premiums have increased over time. How have the subsidy cliff amounts changed over time? Data sets similar to the one analyzed here are available for 2014, 2015, 2016, and 2017. They are not all in identical formats, so a little data wrangling will be necessary before they can be compared.

What explains the geographic patterns?

Why do some states have high subsidy cliffs and others low subsidy cliffs statewide, while others have significant variations across counties?

What policy changes should be made?

What policy changes could be made to affect the size and distribution of subsidy cliffs, and what are the implications if subsidy cliffs grow larger?


Data sources:

Source code repository: https://github.com/davidbody/aca-subsidy-cliff.